About
I am an experienced Engineering Manager with a demonstrated track-record of repeatedly…
Activity
-
#InBetweenJobs When I first arrived in the USA, I often heard people mention they were "in between jobs." When I asked what was next, the typical…
#InBetweenJobs When I first arrived in the USA, I often heard people mention they were "in between jobs." When I asked what was next, the typical…
Liked by Brent Bryan
-
An update on tmu at Google.... After over 14 years at Google, yesterday was my last day. More info on the next steps shortly but it is worth…
An update on tmu at Google.... After over 14 years at Google, yesterday was my last day. More info on the next steps shortly but it is worth…
Liked by Brent Bryan
-
Turns out that being "semi-retired" is kinda boring. I'm joining Vik Gupta, Ariel Bardin, and Leho Nigul at Warner Music Group to help them build a…
Turns out that being "semi-retired" is kinda boring. I'm joining Vik Gupta, Ariel Bardin, and Leho Nigul at Warner Music Group to help them build a…
Liked by Brent Bryan
Experience
Education
Publications
-
Efficiently Computing Minimax Expected-Size Confidence Regions
ICML 2007
Given observed data and a collection of parameterized candidate models, a 1 -- α confidence region in parameter space provides useful insight as to those models which are a good fit to the data, all while keeping the probability of incorrect exclusion below α. With complex models, optimally precise procedures (those with small expected size) are, in practice, difficult to derive; one solution is the Minimax Expected-Size (MES) confidence procedure. The key computational problem of MES is…
Given observed data and a collection of parameterized candidate models, a 1 -- α confidence region in parameter space provides useful insight as to those models which are a good fit to the data, all while keeping the probability of incorrect exclusion below α. With complex models, optimally precise procedures (those with small expected size) are, in practice, difficult to derive; one solution is the Minimax Expected-Size (MES) confidence procedure. The key computational problem of MES is computing a minimax equilibria to a certain zero-sum game. We show that this game is convex with bilinear payoffs, allowing us to apply any convex game solver, including linear programming. Exploiting the sparsity of the matrix, along with using fast linear programming software, allows us to compute approximate minimax expected-size confidence regions orders of magnitude faster than previously published methods. We test these approaches by estimating parameters for a cosmological model.
More activity by Brent
-
White House Presidential Innovation Fellows (PIF) is accepting applications through June 24! As a current fellow, I’ve loved working alongside a…
White House Presidential Innovation Fellows (PIF) is accepting applications through June 24! As a current fellow, I’ve loved working alongside a…
Liked by Brent Bryan
-
After 11.5 years at Google my last day was on Sept 10th. It was an amazing experience but it was really time for a change as it's a very different…
After 11.5 years at Google my last day was on Sept 10th. It was an amazing experience but it was really time for a change as it's a very different…
Liked by Brent Bryan
-
Seeing Chris Monson's reaction to my "leaving Google" post reminded me that I've got at least one more "thank you" and one life lesson to…
Seeing Chris Monson's reaction to my "leaving Google" post reminded me that I've got at least one more "thank you" and one life lesson to…
Liked by Brent Bryan
-
Congratulations to my friends and former coworkers at Pinterest! Excited to see the success of the company and products you have built and proud to…
Congratulations to my friends and former coworkers at Pinterest! Excited to see the success of the company and products you have built and proud to…
Liked by Brent Bryan
-
Going home energized after another great #GHC.
Going home energized after another great #GHC.
Liked by Brent Bryan
-
"When you don't have the vocabulary to describe what you are looking for." Ben Silbermann talking about how Pinterest's visual search is based on our…
"When you don't have the vocabulary to describe what you are looking for." Ben Silbermann talking about how Pinterest's visual search is based on our…
Liked by Brent Bryan
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Brent Bryan in United States
-
Brent Bryan
-
Brent Bryan
-
Brent Bryan
VP at JL Bryan Equipment
-
Brent Bryan
Associate Information Systems Analyst at Department Of California State
13 others named Brent Bryan in United States are on LinkedIn
See others named Brent Bryan